RIP Sony a7s III

jonpais said:
No one in this thread said Chapman had to come out and say anything.
"
Click to expand...

Sony was more than willing to crow about its features to CineD but not lie about a presumed non-existent 16 bit sensor. If it did have such a sensor, there would be no reason to cloak behind secrecy when it clearly prides in promoting them. Mr. Chapman did give his opinion, that it is not a 16 bit sensor.

You have ignored at least two times now, questions posed from Mr. LennyLevy.
 
Dolby if not Netflix does state in this white paper==> Preparation of Papers in Two-Column Format (dolby.com) that quantization distortions such as banding and contouring will be visible in some 10b HDR imagery, and that 12 bit is necessary to decrease the distortions. I prepared a silly little 12 bit ICtCp in 8K HDR for YouTube, hardly worthy of anything that Mr. Doug Jensen could craft in his wildlife rich, rocket filled Florida home but it did get 30,000 views in a short time so there seems some interest in the subject.

 
Tom Roper said:
Dolby if not Netflix does state in this white paper==> Preparation of Papers in Two-Column Format (dolby.com) that quantization distortions such as banding and contouring will be visible in some 10b HDR imagery, and that 12 bit is necessary to decrease the distortions. I prepared a silly little 12 bit ICtCp in 8K HDR for YouTube, hardly worthy of anything that Mr. Doug Jensen could craft in his wildlife rich, rocket filled Florida home but it did get 30,000 views in a short time so there seems some interest in the subject.
Click to expand...

Tom, is your post is in reply to my question to jonpais about Netflix saying 10-bit internal ain't good enough? If so, I'm afraid the Dolby document you provided is way over my head and I don't really have any interest in that kind of detail. I'm staying completely out of this debate you guys are having because I don't really care who is right or wrong. Makes no difference to me because I have no interest in shooting RAW with any of these cameras. But I would like to see some documentation from jonpais where Netflix says 10-bit internal isn't good enough for HDR. I don't know if he's right or not, but I'd like to hear it from the horse's mouth.
 
Ha, ha. Don't drag me in. I'm staying neutral. I just want to see some documentation about this one claim. Should be easy to post if it is true.
 
FWIW, looks like the Netflix Production best practices guide recommends Cine EI, not Inflexible ISO, for the FX3. Makes sense, since it fits better into the professional workflow of the FX6, FX9, Sony Venice and the other 9 or so Sony cameras on the Approved list.
 
Last edited:
You do realize, pilfering documents that you redistribute from your own account is a copyright violation? Why don't you just share the link to the Netflx Partner Help Center?
 
The workflow for recording RAW out from the A7SIII would be semi-similar to cine EI anyway, right? Is the PRR recording truly ISO-agnostic? As in, does it only matter if you're using low or high base ISO, and then your settings in-camera are effectively just an exposure preview? In that case, if you have to record the PRR externally from the FX3 anyway, all that would really keep the A7SIII PRR workflow from Netflix certification might be timecode. Not that I care one way or the other, but it sure seems like you could still get the job done with an A7SIII.

Interesting about the 12-bit requirement for HDR. I imagine this might only be a problem for low-contrast scenes that are recorded using a small number of available values and then have their contrast radically increased
 
ahalpert said:
Interesting about the 12-bit requirement for HDR. I imagine this might only be a problem for low-contrast scenes that are recorded using a small number of available values and then have their contrast radically increased
Click to expand...

Netflix just says 12 bit Dolby Vision, period, is all they will accept for HDR, and that they will produce the other flavors from that, like HDR10, so that's really easy, you just give them a 12 bit 444 master and they do the rest.

It's in the Dolby Vision white paper you'll find the nitty gritty research details, which are shown only for DCI Commercial Cinema. It's entirely possible the findings are not even applicable for viewing criterias other than commercial cinema. The answer to your question is found in the section titled "baseband quantization performance." You are the only one reading this topic that will understand or care. But to summarize, the table is broken down by DeltaE above 3.0 for commercial cinema theater tolerance, DeltaE 1.0 being the visual threshold.

- 10b YC'bC'r is above the DeltaE 3.0 at all luminance betwwen 1 and 10,000 (nit), so no-go for commercial cinema.

- 10b ICtCp is below the DeltaE 3.0 threshold and above DE1.0 at all luminance between 1 and 10,000, so quantization such as banding and contouring is visible but still acceptable for commercial cinema.

- 12b YC'bC'r is below the visual threshold at all luminance greater than 22 nit and still below the commercial tolerance at less than 22 nit.

- 12b ICtCp is below the visual threshold at all luminance 1-10,000 nit.

Preparation of Papers in Two-Column Format (dolby.com)
 
markfpv said:
I'm not gonna reply to any of this nonsense ... as you don't seem to understand anything about the A7sIII and what I've been saying... so I'll just requote how Ahalpert responded... which is exactly right...
Recording in cine EI always records at base iso (high or low). You could record at base ISO on the a7siii and get the same results.
Click to expand...
It is impossible to recreate the Cine EI workflow with the a7s III or a7 IV. When shooting with these cameras, the best practice is stick to base ISO and give the sensor as much light as possible without clipping (ETTR). In the case of S-Log3, around one to two stops ETTR will give the best results: the cleanest possible image with the greatest amount of information from the sensor (for more info about ETTR, see Paul Leeming). By definition, when shooting ETTR, it is not possible to record any more light without clipping the highlights – and starving the sensor of light by underexposing will only result in more noise and less dynamic range in the shadows. It is common practice to employ EI alongside ETTR, as respected filmmakers like Eric Messerschmidt do. To repeat: shooting at base ISO with the a7s III or a7 IV and under- or overexposing each shot in a project will not only not result in improved image quality, it will only make post production more time-consuming.

Lastly, there's a reason why Sony provides a chart illustrating how changing EI redistributes dynamic range which the dozens of videos and articles online ignore completely. A lower EI allocates more dynamic range below middle gray while minimizing noise and a higher EI redistributes more dynamic range above middle gray. So, you can absolutely depart from base ISO when using Cine EI, though Alister Chapman cautions against rating the camera any higher than base EI/ISO because the image will become too noisy. In fact, if you're not taking advantage of the redistribution of dynamic range in Cine EI, you shouldn't expect optimal results.
 
Last edited:
jonpais said:
To repeat: shooting at base ISO with the a7s III or a7 IV and under- or overexposing each shot in a project will not only not result in improved image quality, it will only make post production more time-consuming.
Click to expand...

FALSE. FALSE. FALSE. There is absolutely nothing about overexposing each shot (when shooting slog3cine) by one stop with the a7sIII that makes post more time-consuming. Just as there is also nothing about overexposing each shot - when using FX6, FX9 or FS7, F55 in CineEI mode, that makes post more time consuming.

jonpais said:
In fact, if you're not taking advantage of the redistribution of dynamic range in Cine EI, you shouldn't expect optimal results.
Click to expand...

Oh please, wise master of the CineEI, please show us this "redistribution of dynamic range" you speak of using Sony's charts.
 
You must log in or register to reply here.
Back
Top